Even for empty table, there is still a file (of very small size) exists. So a exactly zero bytes input is retreated as a system error.
On Tue, Oct 17, 2017 at 2:08 PM, 曾耀武 <zeng.ya...@immomo.com> wrote: > 你的表输入数据为空,检查一下你需要计算的原始数据分区是否有数据。 > > 发件人: "li...@fcyun.com" <li...@fcyun.com> > 答复: user <user@kylin.apache.org> > 日期: 2017年10月17日 星期二 下午2:06 > 至: user <user@kylin.apache.org> > 主题: [外部邮件] Map input splits are 0 bytes, something is wrong! > > hi all, > > Have you ever see the error like this? how to fix it? > what are the possible problems to check out? > > > > #7 Step Name: Build Base Cuboid > Duration: 0.01 mins Waiting: 0 seconds > > java.lang.IllegalArgumentException: Map input splits are 0 bytes, something > is wrong! > at > org.apache.kylin.engine.mr.common.AbstractHadoopJob.getTotalMapInputMB(AbstractHadoopJob.java:573) > at org.apache.kylin.engine.mr.steps.CuboidJob.run(CuboidJob.java:134) > at org.apache.kylin.engine.mr.MRUtil.runMRJob(MRUtil.java:102) > at > org.apache.kylin.engine.mr.common.MapReduceExecutable.doWork(MapReduceExecutable.java:123) > at > org.apache.kylin.job.execution.AbstractExecutable.execute(AbstractExecutable.java:124) > at > org.apache.kylin.job.execution.DefaultChainedExecutable.doWork(DefaultChainedExecutable.java:64) > at > org.apache.kylin.job.execution.AbstractExecutable.execute(AbstractExecutable.java:124) > at > org.apache.kylin.job.impl.threadpool.DefaultScheduler$JobRunner.run(DefaultScheduler.java:142) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) > at java.lang.Thread.run(Thread.java:748) > result code:2 > > > thanks a lot ~ > > ------------------------------ > li...@fcyun.com >